Milk exosomes anchored with hydrophilic and zwitterionic motifs enhance mucus permeability for applications in oral gene delivery

Abstract

Exosomes have emerged as a promising tool for the delivery of drugs and genetic materials, owing to their biocompatibility and non-immunogenic nature. However, challenges persist in achieving successful oral delivery due to their susceptibility to degradation in the harsh gastrointestinal (GI) environment and impeded transport across the mucus-epithelium barrier. To overcome these challenges, we have developed high-purity bovine milk exosomes (mExo) as a scalable and efficient oral drug delivery system, which can be customized by incorporating hydrophilic and zwitterionic motifs on their surface. In our study, we observed significantly improved transport rates by 2.5-4.5-fold in native porcine intestinal mucus after the introduction of hydrophilic and zwitterionic surface modifications, as demonstrated by transwell setup and fluorescence recovery after photobleaching (FRAP) analysis. Remarkably, mExo functionalized by a block peptide (BP), consisting of cationic and anionic amino acids arranged in blocks at the two ends, demonstrated superior tolerability in the acidic gastric environment (with a protein recovery rate of 84.8 ± 7.7%) and exhibited a 2.5-fold increase in uptake by intestinal epithelial cells. Furthermore, both mExo and mExo-BP demonstrated successful intracellular delivery of functional siRNA, resulting in up to 65% suppression of the target green fluorescence protein (GFP) gene expression at a low dose of siRNA (5 pmol) without causing significant toxicity. These findings highlight the immense potential of modifying mExo with hydrophilic and zwitterionic motifs for effective oral delivery of siRNA therapies.

figure 1.

schematic of mexo surface modification and enhanced intestinal mucus and epithelium penetration strategies. (a) dspe-peg-azide was anchored on the mexo surface via the insertion of the hydrophobic part of dspe. (b) peptides for surface modification were first attached to dspe-peg-azide through dbco-nhs ester crosslinker. (c) the transport of native mexo through mucus is trapped due to hydrophobic interactions with the hydrophobic domain of mucus and mucus-associated lipids. surface pegylation was introduced to enhance mexo hydrophilicity and thereby its mucin permeability. the mucin-mimicking peptide (mp) has a sequence similar to the hydrophilic pts domain of mucin monomers. zwitterionic motifs, mimicking the virus structure, are also introduced. this includes a small molecule dilauroylphosphatidylcholine (dlpc), a 20 amino acid alternating peptide (ap) with alternating lysine (cationic) and glutamic acid (anionic), and a similar sequenced block peptide (bp) with the cationic and anionic amino acids clustered in blocks at the two ends. the presence of cationic charges enhances mexo-cell communication and uptake. additionally, the stable α-helical structure of bp improves mexo stability in the acidic gastric environment.

figure 2.

the characterization of mexo and the determination of its stability in gastrointestinal (gi) environment. (a) the bca protein concentration of harvested mexo after passing through sec. (b) transmission electron microscopy (tem) images of both native and surface modified mexo. (c) flow cytometry confirmed surface modification of mexo. non-labeled beads were selected for sample selection and intensity control. surface modification of mexo was confirmed using flow cytometry. the surface-modified mexo was dual-labeled and captured by the anti-cd63 coated magnetic beads. fitc-a channel represents the fitc labeled mexo intensity, and apc-a channel represents the cy5 labeled peg or bp intensity. (d) the size change and total protein recovery of native mexo and surface modified mexo were measured after incubating in enzyme-deficient simulated salivary fluid (ssf ph 7, 5 min), simulated gastric fluid (sgf ph 2.2, 2 h), and simulated intestinal fluid (sif ph 7, 2 h). (* vs. corresponding condition before treatment, # vs. mexo after treatment, $ represent statistical difference between mexo-ap and mexo-bp. p< 0.05, n = 3/group)

figure 3.

(a) the relative permeability coefficients (papp) of fitc-labeled native and surface-modified mexo were measured in porcine intestinal mucus using the transwell setup. (b) the relative permeability coefficients (papp) of fitc, mexo, and surface-modified mexo in mucus were compared. (* vs. fitc, # vs. mexo. p<0.05, n=6/group). (c) representative frap curve and parameters of interest. fluorescence intensity before photobleaching (fi), fluorescence after photobleaching (f0), final fluorescence (f∞), characteristic diffusion time ($\mathbf{\tau }1/2$). the fitting equations used to obtain the effective diffusion coefficient (deff) and the equation used to calculate mobile fraction (k) are also shown in the plot. (d) average diffusion pa (dashed line) and k (bar plots) of native and surface-modified mexo in mucus. (* vs. mexo. p< 0.05, n = 3/group)

figure 4.

intestinal epithelial (caco-2) cellular uptake and cytotoxicity. (a) fluorescence colocalization of caco-2 cells treated with exoglow red-stained native and surface-modified mexos. (b) flow cytometry quantified the fluorescence of mexos uptaken by caco-2 cells. (c) the cell viability of caco-2 cells treated with native and surface-modified mexos was analyzed using the mtt assay. (* vs. mexo. p<0.05, n = 6/group)

figure 5.

molecular dynamics (md) simulation of peptides at ph 7 and ph 2.2. (a) the root mean square deviation (rmsd) for both ap and bp are depicted, representing the displacement changes across the entire peptide. (b) the composition of protein secondary structure elements (sse) is shown for each peptide over time and on average. (c) the alphafold-predicted 3d structure of both ap and bp is presented (in the middle) along with their representative structures during md simulation in both ph conditions.

figure 6.

mexo mediated sirna delivery and gene silencing. (a) illustration of sirna-mexo loading and hek293 gfp silencing. (b) fluorescence images of gfp-expressing hek293 cells silenced by lipofectamine-sirna complex, sirna-loaded mexo, mexo-peg, and mexo-bp. (c) normalized gfp fluorescence (rfu) of silenced hek293 cells quantified by flow cytometry. (* vs control, # vs. lipo, p<0.05, n = 4/group)